September 2011 - Dev

Re: [DynareDev] Fwd: RE: identification with partial info
by Michel Juillard 25 Sep '11

25 Sep '11

Hi Bo and Marco, we need first to integrate the code for estimation under partial information, then we will see how to deal with identification. Best Michel On 09/22/2011 05:20 PM, rattoma wrote: > Hi Michel, > > do you think the attached idea is feasible? ( I mean without resorting to > numerical derivatives and using analytic derivations) > best > Marco > > -------- Original Message -------- > Subject: RE: identification with partial info > Date: Thu, 22 Sep 2011 12:57:49 +0100 > From: Bo.Yang(a)surrey.ac.uk > To: marco.ratto(a)jrc.ec.europa.eu > > Hi Marco > > We are hoping to use your identification toolbox under the partial > information setting. > > I think this would involve at least changing the likelihood computation > within the identification toolkit, and perhaps changes at some other > places. > > I wonder if you would provide some guidance on where to focus on. Many > thanks. > > Best > Bo >

1 0

Fwd: Cron <sebastien@karaba> /home/sebastien/dynare/check/matlab/make-check-matlab.sh
by Sébastien Villemot 20 Sep '11

20 Sep '11

At last! -- Sébastien Villemot Researcher in Economics at CEPREMAP & Debian Maintainer http://www.dynare.org/sebastien Phone: +33-1-40-77-49-90 - GPG Key: 4096R/381A7594

1 0

Estimation broken
by stepan＠ithaca.fr 19 Sep '11

19 Sep '11

Dear all, The estimation in dynare/master is broken since (my) commit 322adb92a5ee378ac89610678aeef8e65cebae1e. I am importing new versions of the estimation routines from one of my private branch which have diverged from master. I try to fix this as soon as possible... But it is very likely that I will break again dynare/master in the next weeks (not only the estimation). Best, Stéphane. -- Stéphane Adjemian Université du Maine, Gains et Cepremap Tel(Gains): +33(0)2 43 83 31 35 Tel(Cepremap): +33(0)1 40 77 84 19

4 12

[Trac] #210: block_bytecode test files: the tests should not report an error if Matlab's optimization toolbox is not available
by Trac Dynare 19 Sep '11

19 Sep '11

#210: block_bytecode test files: the tests should not report an error if Matlab's optimization toolbox is not available --------------------------------------+------------------------------------- Reporter: michel | Owner: sebastien Type: bug | Status: new Priority: major | Milestone: Component: Build system / Packaging | Version: Keywords: | --------------------------------------+------------------------------------- One can test for exist('fsolve') in run_test_matlab.m (fsolve is available under Octave) -- Ticket URL: <https://www.dynare.org/trac/ticket/210> Dynare <http://www.dynare.org> The Dynare project

1 2

Re: [DynareDev] dynare.texi
by Sébastien Villemot 14 Sep '11

14 Sep '11

Houtan Bastani <houtanb(a)gmail.com> writes: > Was there an automatic way that you created endlines in dynare.texi? > ie, Instead of having the lines be continuous, you seemed to break > them at some margin. Does it matter to you if I don't do that? I would prefer you to do that, if possible. Using Emacs, it is very easy. The feature is called "filling". Here is the way to do it: - first, configure the variable "fill-column" to 79 - second, when you are editing the .texi file, run "M-x auto-fill-mode"; now Emacs will automatically break lines of more than 79 columns (activating auto-filling can be automatized using hooks, for example by adding "turn-on-auto-fill" to "text-mode-hook") - later, if you modify a paragraph which has already been filled, Emacs will not automatically fill it again; type "M-q" to do ask it to fill again. That's it. Best, -- Sébastien Villemot Researcher in Economics at CEPREMAP & Debian Maintainer http://www.dynare.org/sebastien Phone: +33-1-40-77-49-90 - GPG Key: 4096R/381A7594

2 1

[Trac] #188: reference manual fails to compile under Cygwin
by Trac Dynare 14 Sep '11

14 Sep '11

#188: reference manual fails to compile under Cygwin ---------------------------+------------------------------------------------ Reporter: michel | Owner: Type: bug | Status: new Priority: minor | Milestone: Component: Documentation | Version: 4.2.1 Keywords: | ---------------------------+------------------------------------------------ Lorsqu'on compile le manuel sous Cygwin, on obtient l'erreur dynare.texi:3500: Anchor `estimation' and node `Estimation' map to the same file name. dynare.texi:3500: @anchor command ignored; references to it will not work. dynare.texi:3500: Rename this anchor or use the `--no-split' option. qui est liee au fait que sous Windows les majuscules ne sont pas distinguees des minuscules dans les noms de fichiers. -- Ticket URL: <https://www.dynare.org/trac/ticket/188> Dynare <http://www.dynare.org> The Dynare project

1 1

Re: [DynareDev] [Dynare Git] Dynare branch, master, updated. 1e04330fe897874f41e930b999b6d0ac58c66bca
by stephane.adjemian＠univ-lemans.fr 05 Sep '11

05 Sep '11

Hi Houtan, Yes. I changed the order of the outputs from the kronecker mex files because I don't want to have the error code as the first output argument. This complicates considerably the calls to these routines (we have in some situations to create a lot of temporary variables) and makes the code more difficult to read. Obviously I had to remove the test on the number of output arguments... I tested all these changes and I didn't see any problems. Yes. I introduced inconsistencies in the treatment of the error codes. I think we should change all the mex files... But we have time to discuss this. In my view, the reversed ordering of the outputs is preferable if we have to call a large number of times the mex routines. Best, Stéphane. Houtan Bastani <houtanb(a)gmail.com> writes: > Hi Stephane, > > In commit 560800bcc53b389cc6130bfcc131ff4a5e4bfbc0, you change the order of the error output. What was the reason for this? It makes these mex files inconsistent with all of the other ones. > > Also, why did you remove the check on the size of nlhs in the mex files. This should be there to ensure that nlhs contains an error code as, if nlhs == 0, no error code will be produced (see dynmex.h). Also, in removing the check on nlhs but still assigning to plhs, I think the mex file will crash if there are not enough lhs return values. > > Best, > Houtan > > > On 5 Sep 2011, at 12:54 PM, Stephane Adjemian wrote: > > This is an automated email from the git hooks/post-receive script. It was > generated because a ref change was pushed to the repository containing > the project "Dynare". > > The branch, master has been updated > via 1e04330fe897874f41e930b999b6d0ac58c66bca (commit) > via b923ccd4e8fc3a17b42b112c339d2e57e29deba9 (commit) > via ada3eec81bc6be4bad0a42a13e3b270389a044f1 (commit) > via a792f3f0b36a8a99a010420140fb707c9def94f6 (commit) > via c0faa2ccb8c2dedbfd44908ac7eef461b00d7f0d (commit) > via f3652af93858432e0a2fc1f3c351a5cc715b6baa (commit) > via a73022403128a9496497eaa8381a8ba50a9d1ddd (commit) > via 560800bcc53b389cc6130bfcc131ff4a5e4bfbc0 (commit) > via 9b9fb7df3a71594f15b6cfce6c0307c7515eaac4 (commit) > via 3545be02268275ccbe5253ca5308321948641c17 (commit) > via a2b1ce977f92498b7adf411ae5817be5ec149aa3 (commit) > via d0fe3c8ae449634c476e0b9aa6101281d6898fea (commit) > via 2c8b71ae05cb4cbca821286650607e31bb3c044d (commit) > via ff98f65cf6dda854a645ca90471d1ad4e6aa330d (commit) > via 8982d495399b95acc62ea726a6900d8104c33b14 (commit) > via 31c6f1c998a16e1b313b7a4d3996f2524482664c (commit) > via 237744fe7e05ff3b7ad3358550e4d3883a0b3cd0 (commit) > via 467c60c51e1eec01a12cb2f13ccd9da7a4df1305 (commit) > via 85e0ca379a11e01b92eeb97100b651cdfb611d77 (commit) > via e905ff77c09a0e91da6b336b811279c7c16349d5 (commit) > via fe59061e91419ec8b16af481d038f6f8a976ef04 (commit) > via 5efcf7527d940fae12a385a7577f98048e4cd32b (commit) > via b9c35234e1118ae85eacdf96f605fd41412144f7 (commit) > via 948594c35f386d4d68231616cbaa0fb78d4d5102 (commit) > via 4323653f08af533034faa176e3aca3de168a52f2 (commit) > via 8c49e99ee8c9044520adde5e7c95f448578969b7 (commit) > via 316e16edc2838d434df43bafbc697015377bb154 (commit) > via 8c4e24769b4bb271bbb060c29191b107e5293597 (commit) > via 9b00852e255c30ddca30ffcea63e37c674dfc230 (commit) > via 0c61ddc63a9d12a5e6fedfdabf170c0038a1e918 (commit) > from 8345de43b0489634aab6f0eec6ce0c539a93b0ba (commit) > > Those revisions listed above that are new to this repository have > not appeared on any other notification email; so we list those > revisions in full, below. > -- Stéphane Adjemian Université du Maine, GAINS & CEPREMAP http://www.dynare.org/stepan TEL (Cepremap) +33(0)1 40 77 84 19 TEL (Gains) +33(0)2 43 83 31 35

1 0

EXPECTATION operator with partial information in Dynare
by Sébastien Villemot 01 Sep '11

01 Sep '11

Hi George, Some time ago, we implemented a syntax like that in Dynare: EXPECTATION(full)(SOME_VARIABLE) EXPECTATION(varobs)(SOME_VARIABLE) The idea was that you could, in the same model, variables known with the full information set, and variables known with the partial set. To my knowledge, this syntax is recognized by the preprocessor, auxiliary variables are created, but then it crashes because there are not as many variables than equations. So basically this thing seems broken. Has it ever worked? Have you plans to make it work? Best, -- Sébastien Villemot CEPREMAP - Paris School of Economics Homepage: http://www.dynare.org/sebastien Landline phone: +33 1 40 77 49 90 SIP phone: sebastien.villemot(a)ekiga.net PGP Key: 2069 1DFC C2C9 8C47 9529 84EE 0001 8C22 381A 7594

2 3

Octave sprints
by Sébastien Villemot 01 Sep '11

01 Sep '11

Just for your information, the Octave developpers are organizing a "sprint" on Saturday: basically it is an online gathering of developpers and programmers who collectively work a couple of hours in order to tackle a given objective. The goal for Saturday sprint is to add many new unitary tests. The only knowledge required is MATLAB/Octave programming. For more information: http://www.gnu.org/software/octave/sprint.html -- Sébastien Villemot Researcher in Economics at CEPREMAP & Debian Maintainer http://www.dynare.org/sebastien Phone: +33-1-40-77-49-90 - GPG Key: 4096R/381A7594

1 0